The origins of nominal affixes in Austroasiatic and ... - Roger Blench

Nominal affixes in Austroasiatic and Sino-Tibetan Roger Blench Circulation draft
1. Introduction
The languages of the world characteristically have morphological strategies to classify nouns and signify to
both speakers and hearers aspects of the semantics of the noun. These strategies can be broadly divided into
two widespread categories, noun-classes and numeral classifiers. A third system, characteristic of phyla such
as Indo-European and Afroasiatic, is the sex-gender system, classifies nouns through notional male/female
oppositions, but which is often not informative about semantics. Noun classes are often referred to as
‘genders’ but this is a confusing terminology as it conflates a genuine semantic categorisation with sexgender
systems, which broadly do what they say on the tin. Global overviews such as Aikhenvald (2000)
sometimes conflate these two systems. These categories are not watertight; languages can display aspects of
all these, and erosion of one morphosyntactic category can lead to the partial or complete evolution of
another. For example, the Kana language in the Niger Delta of Niger is historically a Niger-Congo language
descended from languages with rich noun-class systems. However, along with the other members of the
Ogoni group, it has almost entirely lost this aspect of its morphology and developed instead a system of
numeral classifiers (Ikoro 1996). Similarly, Krongo, one of the Kadu languages, a branch of Nilo-Saharan,
appears to have nominal affixes, although these are partly fossilised and unproductive. They have no
semantic assignments, but Krongo has instead adopted or developed a sex-gender system (Reh 1985, 1994;
Blench 2006).
Languages are, broadly speaking, conservative and phyla can be characterised by particular strategies. So the
great majority of Niger-Congo languages have noun-classes or nothing; numeral classifiers or sex-gender
systems rarely develop. Afroasiatic languages have sex-gender throughout the phylum and indeed a highly
conservative morphology. Austroasiatic, Austronesian, Hmong-Mien and Sino-Tibetan languages all have
numeral classifier systems and it is usually considered that noun classes are absent in SE Asia. Concordial
noun classes are widespread, though far from universal, in the Papuan and Australian language areas
(Harvey & Reid 1997), and appear occasionally in New World phyla, in keeping with their great diversity.
This paper asks whether it is actually true that noun classes are absent in SEA Asia. Typical noun class
languages consist of a root and an affix; which can be prefixed or suffixed, and appear as an infix or
circumfix. In rare cases, languages exhibit double-affixing, two separate affixes which alternate according to
distinct rules. Examples of such languages are Bassari on the Togo-Ghana borderland and the Tivoid
languages of SE Nigeria (Greenberg 1977). As part of the erosion of such systems, various types of fusion
can occur, but the original morphemes are usually reconstructible. The affix in principle has a semantic
assignation, and in some languages this is clear, whereas elsewhere the semantics can be opaque. Affixes
frequently alternate; thus singulars can have one or two marked plurals. In Nilo-Saharan it is often
considered that the ‘middle’ is the unmarked term and a singulative and a plurative can be formed from the
root (Dimmendaal 2000). But this is not a necessary requirement of a noun class language; in Niger-Congo
the m- class for mass nouns is always an unpaired class (Greenberg 1963; Blench 1995). Another important
aspect of classic noun-class languages is concord, agreement between the nominal affix and affixes on other
parts of speech in the same clause or sentence. Most typically concord is with adjectives, but verbs,
pronouns and other particles can all enter the picture. Agreement can be alliterative, i.e. the same or
segmentally related material is repeated, or non-alliterative, i.e. the relationship is regular without segmental
copying.
In SE Asian language phyla, word structure is often described as ‘sesquisyllabic’ (Matisoff 1973). The
intended meaning is that words have major and minor syllables, i.e. an iambic structure. The major syllables
are in the stem and the minor syllable a prefix, generally C or CV. Since the –V is often represented
orthographical with a mid-central vowel, it may be that it is not realised segmentally. Minor syllable prefixes
are either lost or optional in many languages, and they seem to be able to change in ways that do not suggest
phonological shift but affix substitution. Austroasiatic and Sino-Tibetan are of this type. In Austronesian
languages, the tendency is for the prefix to have a (C)V format and to be conserved. In other words, once a
prefix has been fused to a stem, it is retained, from Taiwan to New Zealand, as it were. There is an exception
to this, the Daic languages, which are typically CV(C) and except in rare cases no longer retain the minor
syllable. The origin of the Daic system is considered below.
1